Skip to content

sgmurphy/NoiseGate

Repository files navigation

NoiseGate

Filter the noise from bug bounty reports.

A program-aware AI triage assistant that evaluates report quality, scope alignment, and impact — so researchers submit better reports and triagers focus on real vulnerabilities.

Demo

NoiseGate demo

Quickstart

pip install noisegate

1. Configure your LLM provider:

noisegate llm

2. Import a bug bounty program:

noisegate program add acme --url https://hackerone.com/acme

3. Triage a report:

noisegate triage report.md

Each report gets a three-verdict result:

Verdict Meaning Action
PASS In scope, real impact, clear writeup Submit / triage it
WARN Real finding, incomplete writeup Improve the report
FAIL Out of scope or ineligible Don't submit

Usage

noisegate triage report.md                          # default program
noisegate triage report.md --program acme           # specific program
noisegate triage report.md --json                   # machine-readable output
noisegate triage report.md --model openai:gpt-4o    # override model

Exit codes: 0 PASS, 1 WARN, 2 FAIL, 3 error.

noisegate program add acme --url https://hackerone.com/acme
noisegate program add acme --file ./policy.html
noisegate program list
noisegate program show acme
noisegate program use acme
noisegate program remove acme

Supports Anthropic (default), OpenAI, Gemini, and Ollama. Pass --model <provider>:<model-id> to switch.

Documentation

See the full docs for configuration, custom check plugins, and evals.

Development

pip install -e ".[dev]"
playwright install chromium
python -m pytest

License

MIT

About

A policy-aware AI triage assistant that evaluates bug bounty report quality, scope alignment, and impact—so researchers submit better reports and triagers focus on real vulnerabilities.

Topics

Resources

License

Stars

3 stars

Watchers

0 watching

Forks

Contributors

Languages